VR-Training Project

Grundlegendes Praxisprojekt - WS 24/25

Arsine Chinaryan, Anh Hoang Minh Dang, Iris Dizdari, Margarita Khachatryan

Dr. Fabian Scheipl, Dr. Sabine Hoffmann, Daniel Schlichting

Cip-Pool Raum III (Ludwigstrasse 28)

2025-01-13

Outline

1. Overview & Terminology

  • VR-Project
  • Description of data sets
  • Terminology
  • Research questions

2. Data Analysis

  • Cohorts comparison: Similarities and Differences
  • Relationship between Stress indicators and Physiological Measurements
  • Relationship over the rounds
  • Subgroups and Outliers

3. Summary

Outline

1. Overview & Terminology

  • VR-Project
  • Description of data sets
  • Terminology
  • Research questions

2. Data Analysis

3. Summary

Overview of the VR Project

VR-Training: Adapting Virtual Reality Training Applications by Dynamically Adjusting Visual Aspects

Authors: Fabio Genz and Dieter Kranzlmüller

Scenario:
- Users move and place parcels in a VR warehouse using controllers.
- Visual cues: dynamic lighting and color guidance.

Overview of the VR Project

Motivation:
- Static training doesn’t fit everyone: Too easy → boredom, Too hard → anxiety
- Adaptive VR training balances difficulty to improve outcomes.

Adaptive Features:
- Tracks user behavior (head movement) and performance (time, errors).
- Adjusts lighting, object colors, etc., after each training round.

Description of data sets

  • Survey: Cross-sectional study with longitudinal elements
  • 2 Cohorts: Linne and Dame
    • 20 participants in Linne, 80 participants in Dame
    • 9 rounds of training
    • 3 groups of training versions: Adaptive, Non-Adaptive, Control
  • Analyzed data:
    • Demographic data: Gender, Age, Weight, Height
    • Stress indicators
    • Physiological measurement data (PMD)

Terminology

Stress indicators:

  • Cognitive load (1 to 6 with 1 very low and 6 very high)
  • Physical load (1 to 6 with 1 very low and 6 very high)

Physiological Measurement Data (PMD)

  • Heart measurements: Heart rate (HR), Heart rate variability (HRV): SDNN, RMSSD
  • Skin conductance response (SCR) measurements: Skin conductance level (SCL), SCR frequency, SCR amplitude, SCR rising time
  • Eye tracking measurements: Blink rate per minute, Saccade amplitude, Saccade velocity

Research questions

  1. Are there differences between the 2 data cohorts, Linne and Dame?
  2. How are the stress indicators and the physiological measurements related?
  3. Does this correlation change over the rounds?
  4. Are there subgroups within the test subjects that stand out from the rest?

Outline

1. Overview & Terminology

2. Data Analysis

  • Cohorts comparison: Similarities and Differences
    - Relationship between Stress indicators and Physiological Measurements
    - Relationship over the rounds
    - Subgroups and Outliers

3. Summary

Cohorts comparison

Age Patterns Based on Training Versions and Cohort

Cohorts comparison

Gender Proportions Based on Training Versions and Cohort

Cohorts comparison

BMI Across Between Dame and Linne Cohorts

Cohorts comparison

RMSSD Across Training Versions in the Cohorts

Cohorts comparison

Cognitive and Physical Load Across the Cohorts

Research questions

1. Are there differences between the 2 data cohorts, Linne and Dame?

→ Despite demographic and protocol differences, the two cohorts are similar enough to be combined for meaningful analysis


2. How are the stress indicators and the physiological measurements related?

3. Does this correlation change over the rounds?

4. Are there subgroups within the test subjects that stand out from the rest?

Outline

1. Overview & Terminology

2. Data Analysis

- Cohorts comparison: Similarities and Differences
- Relationship between Stress indicators and Physiological Measurements
- Relationship over the rounds
- Subgroups and Outliers

3. Summary

Relationship

  • Spearman correlation
  • Correlation absolute values > 0.1

Research questions

1. Are there differences between the 2 data cohorts, Linne and Dame?

2. How are the stress indicators and the physiological measurements related?

  • Positive correlation: Blink rate
  • Negative correlation: Saccade velocity

⇒ Values of the correlation vary from -0.15 to 0.1, indicating very weak relationship between the stress indicators and the physiological measurements


3. Does this correlation change over the rounds?

4. Are there subgroups within the test subjects that stand out from the rest?

Outline

1. Overview & Terminology

2. Data Analysis

- Cohorts comparison: Similarities and Differences
- Relationship between Stress indicators and Physiological Measurements
- Relationship over the rounds
- Subgroups and Outliers

3. Summary

Relationship over the rounds

Relationship over the rounds

Research questions

1. Are there differences between the 2 data cohorts, Linne and Dame?

2. How are the stress indicators and the physiological measurements related?

3. Does this correlation change over the rounds?

→ The correlation between physiological measurements and stress indicators shows minimal changes across the rounds, correlation values mostly falling within the range of -0.2 to 0.2 and no consistent trend observed.


4. Are there subgroups within the test subjects that stand out from the rest?

Outline

1. Overview & Terminology

2. Data Analysis

- Cohorts comparison: Similarities and Differences
- Relationship between Stress indicators and Physiological Measurements
- Relationship over the rounds
- Subgroups and Outliers

3. Summary

Stress Resiliant Subgroups

Stress Resiliant Subgroups

Stress Resiliant Subgroups

Heart rate variability (HRV), measured by RMSSD, reflects stress management.
The threshold for high HRV was measured using the 0.9 quantile.

Research questions

1. Are there differences between the 2 data cohorts, Linne and Dame?
2. How are the stress indicators and the physiological measurements related?
3. Does this correlation change over the rounds?

4. Are there subgroups within the test subjects that stand out from the rest?

  • Subgroups:
    – Dame Males (Perceived Stress > 4): High cognitive load with low heart rate.
    – Linne Adaptive Males: High cognitive load (Q1) with low to moderate heart rate.
    – Age 20-30: High RMSSD at low stress levels in the combined cohort.
  • Outliers:
    – The only participant in the 51+ age group showed very low cognitive load but had the highest HRV compared to the rest of the cohort.
    – A Subject (NonAdaptive, Linne) with cognitive load higher than 3 and very low heart rate.

Outline

1. Overview & Terminology

2. Data Analysis

3. Summary

Summary

  • The 2 cohorts’ key characteristics are similar enough to be combined for meaningful analysis.
  • The relationships between Stress indicators and PMD are very weak (Correlation range: -0.15 to 0.1)
  • These relationships show minimal change over the rounds with no consistent patterns (Correlation range: -0.2 to 0.2)
  • Certain subgroups within the cohorts show distinct patterns of behavior when comparing PMD with stress indicators, along with a few prominent outliers.

Question

Appendix

Appendix